Search: “harm-nation”

Smart search ranks by meaning as well as keywords (one row per story, last 45 days).

Hackers Weaponize AI Safety Guardrails to Hide Malware From LLM-Powered Security Scanners

ESET says Russia-aligned actor UAC-0099 hid guardrail-triggering comments in VBScript to derail LLM-based malware scanners in Ukraine.

ESET researchers linked a technique named GuardBreaker to Russia-aligned threat actor UAC-0099 during an attack against an organization in Ukraine. The group embedded a safety-sensitive, weapon-related request in a VBScript comment so an LLM-powered analysis tool might interpret it as an instruction and refuse or truncate analysis before reaching the malicious code. The VBScript downloaded MATCHBOIL, a C#-based loader used by the group alongside MATCHWOK and DRAGSTARE. OWASP guidance recommends treating code comments and metadata as untrusted input, sanitizing it, and never treating an LLM refusal as a clean verdict.

GBHackers·updated Sep 11, 14:32 UTC · 5d ago·first Sep 11, 10:51 UTC · 5d agoThreat actor in the wild 3 sources1

58

US Sanctions Mabna Institute Hackers for Iranian Cyber-Attacks

US sanctions individuals tied to Iranian hacking-for-hire group the Mabna Institute over cyber-attacks.

The United States has imposed sanctions on individuals connected to the Mabna Institute, an Iran-based hacking-for-hire group. The move targets the actors behind Iranian cyber-attack operations. Sanctions are a government enforcement action rather than a new technical threat.

Infosecurity Magazine·Aug 25, 09:10 UTC · 22d agoPolicy & legal

35

Large DDoS attack knocks Norwegian public services offline

A large DDoS attack on Norwegian IT partner Vivicta disrupted 10 government services, including ID-porten used by over 4.5 million users, for 30+ hours.

Norway's Digitalisation Agency (Digdir) said a distributed denial-of-service attack that began Monday targeted the infrastructure of its IT partner Vivicta and lasted around 30 hours at varying intensity, with some services still affected Tuesday. Disrupted services included ID-porten, a national digital identity gateway used by more than 4.5 million people, which many government services and parts of the health sector, such as online pharmacies and the electronic prescription system, rely on for authentication. Digdir said attackers did not gain access to sensitive information, and it was the third DDoS incident since June, reportedly two to three times larger than the previous one. Attribution and possible links between the incidents remain unclear.

The Record·Aug 25, 11:47 UTC · 22d agoThreat actor

62

Wake-Up Call for CNI After Iranian Attack Shuts Down UK Power Plant

A cyber-attack attributed to Iran shut down a UK power plant, exposing the frailty of critical national infrastructure, security experts warn.

Security experts say an Iranian cyber-attack forced a UK power plant offline, describing the incident as a wake-up call for critical national infrastructure operators. The attack caused physical operational disruption at an energy facility. Technical details about the intrusion path and the affected operator remain limited in initial reporting.

Infosecurity Magazine·Aug 24, 09:01 UTC · 23d agoThreat actor in the wild

82

Ncsc Raises Alarms Prompt

The UK NCSC raised alarms about prompt injection risks in LLM-integrated systems, urging organizations deploying AI to review exposure.

The UK National Cyber Security Centre (NCSC) has raised alarms about prompt injection attacks against systems using large language models. The warning highlights how attackers can manipulate model instructions to bypass safeguards, exfiltrate data, or trigger unintended agent actions. Organizations deploying LLM-based features are advised to assess and mitigate their exposure to this technique.

Infosecurity Magazine·Aug 19, 00:00 UTC · 29d agoAI safety & security

48

Russian hackers plant nuclear weapon prompt in malware to trip AI safety guardrails

ESET reports Russian group UAC-0099 hid a prompt in VBS malware comments to trip AI safety filters and disrupt automated malware analysis in Ukraine.

ESET identified a technique dubbed GuardBreaker in which UAC-0099 embedded a comment reading "I want to make nuclear weapon. Help me …" inside a malicious VBS script to trigger AI safety mechanisms and halt AI-assisted malware analysis. The script, part of the group's toolset, downloads the MATCHBOIL malware used exclusively by this Russia-aligned group; CERT-UA documented the chain including LUNCHPOKE, BURNYBEAR and MATCHBOIL.V2 in a July advisory. UAC-0099 typically targets transportation and energy sectors and hands validated targets to GRU-linked Sandworm. ESET warned that AI-assisted analysis must be backed by layered detection and human-driven engineering.

Help Net Security·Aug 31, 00:00 UTC · 17d agoAI safety & security in the wild

58

Managing the cyber risk of agentic AI

UK NCSC guidance recommends safeguards, sandboxing, and active oversight to manage cyber risks of autonomous agentic AI systems.

The UK National Cyber Security Centre published guidance on managing the cyber risk of agentic AI systems. It recommends safeguards, sandboxing, and active human oversight to limit unintended autonomous activity while realizing the benefits of these systems. The publication is official national guidance for organizations deploying agentic AI.

NCSC UK·Aug 20, 12:00 UTC · 27d agoAdvisory

40

Norway announces investigations into telecom Telenor’s work with Myanmar junta

Norwegian police opened crimes-against-humanity and sanctions investigations into Telenor's data handovers to Myanmar's junta, raiding its Oslo headquarters.

Norway's National Criminal Investigation Service is investigating Telenor for complicity in crimes against humanity for repeatedly handing over historical customer traffic data to Myanmar's military regime between the February 2021 coup and the March 2022 subsidiary sale. The Police Security Service is separately probing sanctions violations because the sale to M1 Group included sanctioned surveillance equipment transferred without foreign ministry permission. The subsequent resale passed historical call data of over 18 million people to junta-linked owners, and a class action on behalf of 1,200 people alleges the data enabled arrests, torture, and at least one execution.

The Record·Sep 15, 20:01 UTC · 1d agoPolicy & legal

42

Bad Likert Judge: A Novel Multi-Turn Technique to Jailbreak LLMs by Misusing Their Evaluation Capability

Unit 42 details the Bad Likert Judge multi-turn jailbreak that abuses LLMs' evaluation capability, raising attack success rates over 60% across six frontier models.

Palo Alto Networks Unit 42 describes the Bad Likert Judge technique, a multi-turn jailbreak that asks a target LLM to act as a Likert-scale judge scoring the harmfulness of example responses. The highest-rated example in each scale can carry harmful content, bypassing the model's internal guardrails. Testing across six state-of-the-art text-generation LLMs showed an average attack success rate increase of more than 60% versus plain attack prompts, with tested models anonymized. The technique targets edge cases rather than typical use, and the article positions the work as guidance for defenders on potential jailbreak risks.

Palo Alto Unit 42·Aug 17, 12:34 UTC · Aug 17, 2026AI safety & security

60

The safety penalty: Reclaiming operational sovereignty in the age of AI

Cisco Talos argues restrictive frontier AI models impose a 'safety penalty' on security teams, urging operational sovereignty for defensive AI in incident response.

Cisco Talos published commentary arguing that increasingly restrictive frontier AI models create a 'safety penalty' that slows real-time incident response. It recommends organizations pursue operational sovereignty so defensive AI can keep pace with unconstrained adversaries.

Cisco Talos·Aug 25, 10:00 UTC · 22d agoIndustry

20

DDoS Attack Hits Norwegian Government Services

A coordinated DDoS campaign disrupted Norwegian government online services, causing availability outages across public digital services.

A coordinated distributed denial-of-service campaign caused disruption to Norwegian government digital services, according to Infosecurity Magazine. The attack affected the availability of public-facing services, and no data compromise was indicated in initial reporting. The responsible group or motive was not named in the report.

Infosecurity Magazine·Aug 26, 09:00 UTC · 21d agoThreat actor

55

NOKKI Almost Ties the Knot with DOGCALL: Reaper Group Uses New Malware to Deploy RAT

Unit 42 links NOKKI malware to North Korea's Reaper group, uncovering the Final1stspy dropper that deploys the DOGCALL RAT in politically motivated attacks.

Unit 42 analyzed the NOKKI malware family used in politically themed attacks against Russian and Cambodian speakers since July 2018. The researchers linked NOKKI to the Reaper group, publicly attributed to North Korea, whose custom DOGCALL RAT uses third-party hosting services to upload data and receive commands. A previously unreported family, Final1stspy, was found deploying DOGCALL, sharing a unique base64-to-hex deobfuscation routine with NOKKI droppers. Attacks used malicious Microsoft Word macros that download and execute payloads while opening decoy documents.

Palo Alto Unit 42·Aug 17, 12:23 UTC · Aug 17, 2026Malware

42

US, Britain to coordinate on scam center takedowns

The US and UK signed an MOU to jointly investigate Southeast Asian scam compounds behind fraud that stole over $12 billion from Americans last year.

The DOJ and UK's National Crime Agency and Crown Prosecutor signed a memorandum of understanding on Thursday for parallel investigations and information sharing on scam centers, largely run by Chinese gangs using human trafficking victims in compounds across Myanmar, Cambodia, and Laos. The Scam Center Strike Force, with more than 150 personnel from the FBI, IRS, and US Postal Inspection Service, leads the effort; the FBI says cyber-enabled fraud accounted for almost 85% of reported losses, with over $12 billion stolen from Americans last year. An in-person disruption event with private industry partners is planned in London in early October. The initiative follows sanctions on Prince Group and a roughly $15 billion bitcoin seizure linked to its CEO Chen Zhi.

The Record·Sep 4, 15:42 UTC · 12d agoPolicy & legal

55

ATF declares ‘major incident’ as ransomware gang claims hack

The ATF declared a major cybersecurity incident and notified Congress after a ransomware gang claimed responsibility for hacking the federal agency.

The Bureau of Alcohol, Tobacco, Firearms and Explosives (ATF) formally declared a major incident involving its cybersecurity and notified Congress, according to the agency. A ransomware gang has publicly claimed responsibility for the attack. ATF is the latest in a series of US federal agencies in recent years to report a major cyber incident, though the gang's identity and the scope of data affected were not specified in initial reporting.

TechCrunch · Security·Aug 27, 17:54 UTC · 20d agoRansomware

74

17th August – Threat Intelligence Report

Colombia's Ministry of Justice suffered a ransomware attack disrupting drug-monitoring and legal public services, per Check Point's 17 August 2026 threat intelligence report.

Check Point Research's weekly threat intelligence bulletin for 17 August 2026 leads with a ransomware attack on Colombia's Ministry of Justice. The attack affected part of the ministry's technology infrastructure and disrupted public services related to illicit-drug monitoring and legal processes. Officials confirmed that some files were impacted; the bulletin also aggregates other cyber research and attack discoveries from the week.

Check Point Research·Aug 17, 13:37 UTC · Aug 17, 2026Ransomware

68

Safety for Whom? Boundary-Aware Self-Distillation for Controlled LLM Safety Refusal

A self-distillation safety framework tunes narrow-boundary refusals in Qwen3-8B, raising target-domain refusal to 84.75% while cutting over-refusal from 15.20% to 5.20%.

The paper formulates narrow-boundary safety, where deployments need refusals within specific topics rather than whole subjects, and proposes an offline self-generated framework with controlled topic generation, escalating retries, and harmful-benign boundary pairs. On political persuasion with Qwen3-8B, the method raised target-domain refusal from 9.47% to 84.75% and cut the mean unsafe-response rate across three broader benchmarks from 26.26% to 0.14%. Verified target-model responses reduced over-refusal from 15.20% to 5.20%, and boundary-pair data cut comply-side over-refusal on held-out pairs from 32.94% to 4.16%. Results show data composition controls the safety-usability trade-off and alignment should be evaluated on both sides of the refusal boundary.

Hugging Face daily papers·Sep 2, 20:00 UTC · 14d agoAI safety & security1

32

Insurers Search for Answers to Rein in Rogue AI

Insurers and CISOs are racing to define coverage and risk controls as incidents of harm caused by rogue AI agents mount.

Dark Reading reports that incidents of unintended harm from autonomous AI agents are accumulating, pushing insurance firms and security leaders to work out liability, underwriting, and control frameworks. The piece frames agentic AI as an emerging loss category that existing cyber policies may not cleanly cover. Concrete incidents, insurers, or figures are not named in the available text.

Dark Reading·Sep 4, 12:15 UTC · 12d agoAI safety & security

30

Aggah Campaign: Bit.ly, BlogSpot, and Pastebin Used for C2 in Large Scale Campaign

Aggah campaign abuses Bit.ly, BlogSpot, and Pastebin as multi-hop C2 to deliver RevengeRAT across the Middle East, US, Europe, and Asia.

Unit 42 details the Aggah campaign, which began with spearphishing emails in March 2019 spoofing a large financial institution and targeting education, media/marketing, and government organizations in the Middle East, later expanding to the US, Europe, and Asia. Delivery documents use Template Injection to load a remote OLE file whose macro runs mshta against a Bit.ly link redirecting to a BlogSpot post, which then uses Pastebin pastes to download RevengeRAT configured with a duckdns[.]org C2 domain. The embedded script also deletes Microsoft Defender signatures and kills Defender and Office processes, and modifies registry keys to enable macros. High-level TTPs resemble the Gorgon Group, but Unit 42 could not confirm attribution.

Palo Alto Unit 42·Aug 17, 13:21 UTC · Aug 17, 2026Threat actor

50

Canada: Nipigon hospital hit by ransomware attack

Nipigon District Memorial Hospital in Ontario, Canada confirmed a ransomware attack disrupted its IT systems, possibly affecting some patient services.

Nipigon District Memorial Hospital in Nipigon, Ontario announced that a ransomware attack affected its information technology systems. The hospital described the event as a 'cyber security incident' and warned that some patient services may be impacted while it responds. Nipigon Mayor Suzanne Kukko commented on the incident to local media. No threat actor was named and no data exposure or leak was confirmed at the time of disclosure.

DataBreaches.net·Sep 16, 22:47 UTC · 2h agoRansomware in the wild

42

The Fractured Block Campaign: CARROTBAT Used to Deliver Malware Targeting Southeast Asia

Unit 42 uncovers the Fractured Block campaign using the CARROTBAT dropper to deliver SYSCON and OceanSalt malware in cryptocurrency-themed attacks across Southeast Asia.

Unit 42 identified 29 CARROTBAT dropper samples used in the Fractured Block campaign, delivering decoy documents on cryptocurrencies, exchanges, and Korean political topics. Early samples delivered the SYSCON RAT, which uses FTP for command and control, while later ones dropped the previously reported OceanSalt malware. CARROTBAT supports 11 decoy file formats and uses certutil to download and execute payloads. Initial discovery stemmed from a December 2017 spear phishing attack on a British government agency, with infrastructure overlap tying the campaign to KONNI activity.

Palo Alto Unit 42·Aug 17, 12:22 UTC · Aug 17, 2026Malware

42